MediMind
Personal Digital Health Assistant
MediMind is a personal digital healthcare assistant powered by AI. It is designed for the older generation that are most affected in terms of lack of health literacy and its negative correlation with overall care, self-management, and health outcomes.
Older generation are most affected in terms of lack of health literacy and its negative correlation with overall care, self-management, and health outcomes. The specific areas in need of support include accessing health information, fully understanding the information from these sources, and being more proactive about the self-management of their care. This support is required because being able to readily access and understand more information about your health, i.e. attaining a higher health literacy level, increases preventative measures in issues with overall health.
Literature Review
Research has shown that certain demographic populations are most affected in terms of lack of health literacy and its negative correlation with overall care. Specifically, those with low socioeconomic status (limited education, resources, and technological skills) and older people are more prone to lack of health literacy.
Unstructured Observations
Patients are often presented with complicated terminology in physical copies of brochures, handouts, and aftercare guides following medical appointments and procedures. The information can often be difficult to digest and understand.
Surveys & Questionnaires
Among interview participants, patients scored an average health literacy score of 3 out of 5. Patients prefer that medical explanations are done using layman’s terms as medical jargon can cause confusion..
Semi-Structured Interviews
Currently, there is a disconnect between the information a doctor gives to a patient and understanding that information for future application. Unfamiliarity with modern technology can cause hesitation to clear any uncertainties via internet searches for the older generation.
Anita
Age: 68
Health condition: Diabetes, Knee surgery
Task: Find out if new symptom is a side effect of her new medication
Joe
Age: 78
Health condition: Occasional back pain
Task: Schedule visit with doctor regarding back pain
Barbara
Age: 73
Health condition: Arthritis, High blood pressure
Task: Find out more information about arthritis diagnosis
Functional
Educating the patient: answer questions about symptoms, issues, etc. & display diagrams that aid in comprehension of the answer
Making appointments easier
Non-functional
Privacy setting/idle mode: certain things aren’t displayed when privacy setting is on
Promote health literacy: user had easy access to information.
Task Scenario
Imagine you are an elderly person who recently had a leg surgery. You are currently taking three different medications as a part of your aftercare regimen. Due to supply issues, your doctor had to recently switch one of these medications. This morning you woke up with throat pain and mild coughing. Log into your health device and use the search feature to figure out if this is a side effect of the new medication.
The prescription search feature showed you that only the mild coughing was a side effect of the new medication. Now use the search feature to find out more information about your throat pain and watch the narrated video it displays.
After finding out more information, you want to meet with your doctor. Use the schedule a visit feature on the home page to call your doctor’s receptionist.
MediMind device in idle mode serves as a digital picture frame.
Patient asks out about a symptom, prescription etc.
Device fetches patient’s information from patient portal & AI feature prompts the user to interact with it.
Narrated video with explanations appear.
User has a personalized conversation with the device about their questions.
User schedules appointment with a medical professional with their provided contact details.
Participants
The workshop was done with 3 participants in their mid-to-late 70’s in their home and 3 college aged participants in studio
Toolkit
This research toolkit was designed to hand out to potential users. Our team would set up a scenario for participants to design for, and interact with to show us their vision of a literary health device to use in an everyday setting.
Findings
A separate device, solely for health information, was preferred by 5 out of 6 participants.
Younger participants included a secondary device in addition to the primary device that was either a wearable or a separate device
Older participants preferred one singular device that was less likely to be misplaced and for ease of use
Low Fidelity Prototype
Evaluation Goals
Navigation
- Can older people use the device to easily find/digest their medical information?
- Is utilizing the AI difficult? How could it be made more intuitive to use for people who are not familiar with AI/Virtual Assistant technology.Appeal
- Is our device appealing to the older generation?
- Would our device be preferred to using a phone or laptop?
- Is our device useful long-term?
Methods
Ask users to “Think Aloud” while partaking in user testing
Retrospective Interview: post-test interview to clarify events that occurred during system use
Metrics
Time: time for users to complete a task successfully
Errors: number of taps that deviate from correct task path
Satisfaction: Survey at end of testing to gauge satisfaction
Measures
Screen recording of UI
Video recording users
Taking notes on thinking aloud data, asking questions
Results
Average 4/5 on satisfaction survey regarding intuitiveness and visual appeal
Average 3.35 seconds to tap after task command is given
No errors made, slight confusion on schedule a visit page
Next Steps
Make voice feature more prominent
Add walkthrough tutorial for first time users
Continue physical model iteration